Round logo for The Computer Church -- a collection of rare computers and vintage documents that catalog the history of computing

The Computer Church
Home       Visit Us       Contact Us       About Us       Volunteers       Donate      
 


Overview       Card Codes       Occupation Codes       Birthplace Codes       Why Take a Census?

1890 Census & Two Brilliant Ideas

image of the <b>The Tablulator</b><br />Part of the<br />Electric Tabulating System
The Tablulator
Part of the
Electric Tabulating System

It was 1889 and the Federal government was facing a HUGE problem. The Department of Commerce had just completed the 1880 census and it had taken a little over 8 years to count the people of the US. Estimates were that the upcoming 1890 census would take MORE than 10 years to complete. The government would be starting the 1900 census BEFORE they had completed the 1890 census!

So in 1889, the Secretary of Commerce held a competition to find a new and faster way to count the people of the United States. The competitors had to count a sample from the previous census (using the raw data from four districts in St. Louis).


1. Electricity

Herman Hollerith had the idea of using electricity to make the tabulation faster. He had been working on an electricity based machine called the "Electric Tabulating System" and he was one of the three finalists in the competition. Hollerith's system easily won the competition and was awarded the contract to count our nation for the 1890 census.

The competition tested two main areas:
Transcribing: the time needed to move the handwritten answers from the individual census questionnaires to the cards.
Tabulation: the time needed to use the cards to actually conduct the count; how many people, how many males, how many can read, etc.

TRANSCRIBING PHASE TABULATING PHASE
Hollerith's method 72 hours 27 minutes Hollerith's method 5 hours 28 minutes
Pidgin's method 110 hours 56 minutes Pidgin's method 44 hours 41 minutes
Hunt's method 144 hours 25 minutes Hunt's method 55 hours 22 minutes

Hollerith won both phases of the competition but the use of electricity gave his system an astounding advantage when the actual tabulation was being done. His system was 8 times faster than his nearest competitor.

Nowadays, we are used to electrical gadgets but in the 1880's, Hollerith's idea was unique. Hollerith was the first person to use electricity to compute numbers. It was so unusual that, in 1890 the Franklin Institute, in Philadelphia, awarded Hollerith the Cresson medal and said of Hollerith, "…the inventor is deserving of the greatest commendation for this useful and novel application of electricity…"


2. Punch Cards

Hollerith had a second brilliant idea. He would punch holes in a card to store the information. Each card would represent a person and the holes would represent characteristic of that person (age, gender, occupation, etc.). The other competitors used cards to represent a person also. However, one wrote on different colored cards for different qualities and the other used different colored inks on his cards. Neither approach was as efficient as Hollerith's method. In fact, Hollerith's punch cards were such an efficient storage method that for about half a century (roughly 1920 to 1970), the cards "...held nearly all of the world's known information...".1



Hollerith Card: This is exactly how an original Hollerith card from the 1890 census would have looked.  The card has no markings other than a code number on the front and back.
(Hollerith Card: This is exactly how an original Hollerith card from the 1890 census would have looked. The card had no markings other than a code number on the front and back.)


Hollerith Card: This is a reader card which showed the coding for each potential punched hole.
(Hollerith Card: This is a reader card which showed the coding for each potential punched hole.
Do you want to know what those codes stood for? See the Card Codes.)


An article that appeared in the The Electrical Engineer pointed out that "…some operators handled in a single day schedules representing as many as 50,000 persons (9,200 families); and that the entire census (comprising 12,690,151 families) was counted in a period of little more than 6 weeks."2

That same article reported astounding numbers. According to the reporter, the 81 tabulating clerks working in the Department of Commerce, had "handled 556,346 cards that day, an average of 6,686 each. The "roll of honor" comprising the best individual daily records of successive weeks, ranges from 9,230 (with 4 readings) up to 13,356 (with 9 readings) with the highest individual record for a single day standing at 19,071 (with 9 readings)".

These cards were commonly known as Hollerith Cards until the 1920's when they became known as IBM cards. Although the cards were amazingly simple (as many great ideas are), they became the foundation for IBM's success. Even into the mid-1950's, the Hollerith punch card made up 30 percent of of IBM's annual bottom line revenue. 3

In the very beginning, as Hollerith was still developing the Electric Tabulating System, he used something we still see today, a train conductor's holepunch, to punch holes in the cards. Hollerith described the process; "The holes may be punched with any ordinary ticket punch, cutting a round hole 3/16 of an inch in diameter." However, as he punched more and more cards, it became obvious that the train conductor's holepunch was a poor choice because the user's hand would quickly cramp up. So Hollerith developed a pantagraph for the 1890 census. In 1901, he would develop a more efficient machine for punching holes in the cards, the Type 001 Keypunch.

Hollerith's Electric Tabulating Machine Company eventually became IBM.





1 IBM 100: Icons of Progress, http://www-03.ibm.com/ibm/history/ibm100/us/en/icons/punchcard/.

2The Electrical Engineer, November 11, 1891.

3 IBM 100: Icons of Progress, http://www-03.ibm.com/ibm/history/ibm100/us/en/icons/punchcard/.